Weight of evidence (WOE) and Information value (IV).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Weight of evidence (WOE) and Information value (IV)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Rating</th>\n",
       "      <th>Income1</th>\n",
       "      <th>Income2</th>\n",
       "      <th>Default</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>B</td>\n",
       "      <td>95.136781</td>\n",
       "      <td>96.819795</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>C</td>\n",
       "      <td>109.999996</td>\n",
       "      <td>103.394892</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A</td>\n",
       "      <td>96.909480</td>\n",
       "      <td>89.676331</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B</td>\n",
       "      <td>99.478898</td>\n",
       "      <td>101.700365</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>D</td>\n",
       "      <td>83.762867</td>\n",
       "      <td>112.293057</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Rating     Income1     Income2  Default\n",
       "0      B   95.136781   96.819795        0\n",
       "1      C  109.999996  103.394892        1\n",
       "2      A   96.909480   89.676331        1\n",
       "3      B   99.478898  101.700365        0\n",
       "4      D   83.762867  112.293057        1"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame({'Rating': np.random.choice(['A', 'B', 'C', 'D'], size=(100)),\n",
    "                   'Income1': 100 + 10 * np.random.randn(100),\n",
    "                   'Income2': 100 + 10 * np.random.randn(100),\n",
    "                   'Default': np.random.choice([0,1], size=(100))})\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>Default</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Rating</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A</th>\n",
       "      <td>0.368421</td>\n",
       "      <td>0.279070</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>B</th>\n",
       "      <td>0.298246</td>\n",
       "      <td>0.232558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C</th>\n",
       "      <td>0.157895</td>\n",
       "      <td>0.162791</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D</th>\n",
       "      <td>0.175439</td>\n",
       "      <td>0.325581</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Default         0         1\n",
       "Rating                     \n",
       "A        0.368421  0.279070\n",
       "B        0.298246  0.232558\n",
       "C        0.157895  0.162791\n",
       "D        0.175439  0.325581"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature,target = 'Rating', 'Default'\n",
    "pd.crosstab(df[feature], df[target], normalize='columns')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>Default</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>woe</th>\n",
       "      <th>iv</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Rating</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A</th>\n",
       "      <td>0.368421</td>\n",
       "      <td>0.279070</td>\n",
       "      <td>-0.277765</td>\n",
       "      <td>0.134146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>B</th>\n",
       "      <td>0.298246</td>\n",
       "      <td>0.232558</td>\n",
       "      <td>-0.248777</td>\n",
       "      <td>0.134146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C</th>\n",
       "      <td>0.157895</td>\n",
       "      <td>0.162791</td>\n",
       "      <td>0.030537</td>\n",
       "      <td>0.134146</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D</th>\n",
       "      <td>0.175439</td>\n",
       "      <td>0.325581</td>\n",
       "      <td>0.618323</td>\n",
       "      <td>0.134146</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Default         0         1       woe        iv\n",
       "Rating                                         \n",
       "A        0.368421  0.279070 -0.277765  0.134146\n",
       "B        0.298246  0.232558 -0.248777  0.134146\n",
       "C        0.157895  0.162791  0.030537  0.134146\n",
       "D        0.175439  0.325581  0.618323  0.134146"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_woe_iv = (pd.crosstab(df[feature], df[target], normalize='columns')\n",
    "             .assign(woe=lambda dfx: np.log(dfx[1] / dfx[0]))\n",
    "             .assign(iv=lambda dfx: np.sum(dfx['woe']*(dfx[1]-dfx[0]))))\n",
    "\n",
    "df_woe_iv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Calculate information value\n",
    "def calc_iv(df, list_features, target):\n",
    "    \n",
    "    df_ivs = pd.DataFrame()\n",
    "    \n",
    "    for feature in list_features:\n",
    "\n",
    "        lst = []\n",
    "\n",
    "        for i in range(df[feature].nunique()):\n",
    "            val = list(df[feature].unique())[i]\n",
    "            lst.append([feature, val, df[df[feature] == val].count()[feature], df[(df[feature] == val) & (df[target] == 1)].count()[feature]])\n",
    "\n",
    "        data = pd.DataFrame(lst, columns=['Variable', 'Value', 'All', 'Bad'])\n",
    "        data = data[data['Bad'] > 0]\n",
    "\n",
    "        data['Share'] = data['All'] / data['All'].sum()\n",
    "        data['Bad Rate'] = data['Bad'] / data['All']\n",
    "        data['Distribution Good'] = (data['All'] - data['Bad']) / (data['All'].sum() - data['Bad'].sum())\n",
    "        data['Distribution Bad'] = data['Bad'] / data['Bad'].sum()\n",
    "        data['WoE'] = np.log(data['Distribution Good'] / data['Distribution Bad'])\n",
    "        data['IV'] = (data['WoE'] * (data['Distribution Good'] - data['Distribution Bad'])).sum()\n",
    "\n",
    "        data = data.sort_values(by=['Variable', 'Value'], ascending=True)\n",
    "\n",
    "        iv = data['IV'].values[0]\n",
    "        s = pd.Series([feature, iv])\n",
    "        df_ivs = df_ivs.append(s, ignore_index=True)\n",
    "    \n",
    "    df_ivs.columns = ['Feature', 'IV']\n",
    "    df_ivs = df_ivs.sort_values(by='IV', ascending=False)\n",
    "    return df_ivs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature</th>\n",
       "      <th>IV</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Income1</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Income2</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Feature   IV\n",
       "0  Income1  0.0\n",
       "1  Income2  0.0"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "calc_iv(df, ['Income1', 'Income2'], 'Default')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}